This change adds deadlock protection to all places where we lock
one thread while waiting for the other to do something. If we
detect that the other thread is going to block, we abort the
operation. This could cause unexpected problems, such as painting
errors, text input errors, or even crashes, but the alternative is
a guaranteed deadlock.
Task-number: QTBUG-41369
Change-Id: I2627a955cfafc4bce54eb9d0d38e19b768b06956
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
Android starts asking lots of questions when the software keyboard
pops up, so let's wait until the application is initialized. This
works around a deadlock on startup.
Task-number: QTBUG-41369
Change-Id: I1c79e32d08c7cc11748ec55efbff3bc25e40f4b2
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
Should not really happen, but since we don't store the focus
object given to us, we should do a check.
A crash was seen from this when running the "Application"
example for widgets.
Change-Id: I9c4121766d7028a4eceede7d7b15c8c53d34e16e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
The reason is that the sender is sometimes 'NULL', so
we cannot rely on it. But the current test with our
prefix should suffice.
Change-Id: Ie58bf062cbade08feda622bda753d63e1d811a8d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Method "- (id)targetForAction:(SEL)action withSender:(id)sender" is
only available from iOS7. So change implementation to use
whats available on iOS 6.
Change-Id: I4e21495073364e83ef396dfab47a7ea2a23bbead
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
The QClipboard documentation states that on OS X it will emit
dataChanged() when activating the application, if the system clipboard
had changed. It wasn't doing this in Qt5.
Task-number: QTBUG-34941
Change-Id: I7f34e757876757691f0a6c94dd2ae76a60146291
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
QDockWidget and QToolBar set the Qt::BypassWindowManagerHint window flag
when unplugging and clear it in the endDrag() methods. This does not have
any effect since the attribute is not taken into account in
QXcbWindow::setWindowFlags(). Change the method to always set the attributes,
which should also make it possible to set/clear Qt::WindowTransparentForInput.
Task-number: QTBUG-41189
Task-number: QTBUG-38964
Change-Id: Id9eddc642489d18f44c7597f8fc1a1df71971306
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
QXcbKeyboard::keysymToQtKey() may return a value of 0 for keysyms
that are not supported by Qt (e.g. for XK_ISO_Next_Group).
It is then translated to an empty QKeySequence which in turn is
detected as a partial match for any shortcut.
The behavior of QShortcutMap::nextState() becomes broken because
it sets current state to QKeySequence::PartialMatch while there is
no match at all.
Task-number: QTCREATORBUG-9589
Change-Id: I1e2a4511a876dfa418db9906d10382255a2e4d62
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
Reviewed-by: Sergey Belyashov <Sergey.Belyashov@gmail.com>
Port change 3716a76704273fdbe5ad4ec978438daeda606c26 (Qt 4)
to Qt 5. Enforce the creation of the internal window
and registering of timers in the event dispatcher constructor
for GUI applications instead of delaying it to processEvents()
is called. Move the call to virtual wakeUp() out of
createInternalHwnd().
Task-number: QTBUG-40881
Change-Id: I82a4748897da140a39feff882c75ad5ac6155148
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Previously, the behavior was different depending on if the
contentDescription was calculated as a result of an event, or if it was
calculated as a result of hierarchy traversal.
Refactor the functionality into one single function that will be used in
both scenarios.
'contentDescription' will now receive its value from one of the following
sources, listed in prioritised order (QAI == QAccessibleInterface):
1. QAI::text(QAccessible::Name)
2. QAI::text(QAccessible::Description)
3. QAI::text(QAccessible::Value)
4. QAI::valueInterface()->currentValue()
Change-Id: I2e4958a1e95b5f20d01da37c23ecbc09842360bc
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@theqtcompany.com>
This disables animations for windows that are
completely obscured by other windows.
On examples/quick/animation, obscured CPU usage goes
from 10% to 1%.
Change-Id: I9945431e6387e406e2064c08d5aa01d5d96ef602
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
QClipboard is documented to take ownership over
the mime data set with "setMimeData" and the value
returned by "mimeData". So we need to implement
this to avoid memory leaks.
Change-Id: Ieb3a17368ed3a698c29a7f92c8ee87a0cca86b46
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Two-part fix: QCocoaMenu::syncMenuItem, when selecting the "old" menu,
if an item was merged, 'applicationMenu' was always selected, but this
is wrong for any item with a role >= CutRole (such an item still can
be "merged", but it's not in the application menu).
QCocoaMenuItem::sync - item can be merged with itself: after item's
role detected, the search for an item to merge with can find exactly
the same item we've just detected the role for (since a data-member is
modified) - try to avoid this.
Task-number: QTBUG-39934
Change-Id: Ibe1df9e92973380652101143067e14922afdfb9e
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
On iOS we do deleteLater() on the mime data instead of deleting
it directly, in case the application should happen to use the
pointer again directly after setting it on the clipboard.
Technically it would be a bug in the application, but using
deleteLater() is safer and it's better to be consistent with
iOS so that a buggy application crashes in the same places in
both.
Change-Id: I2996d6c7816a2f83615a43609f5be207aaa72c86
Task-number: QTBUG-41853
Reviewed-by: BogDan Vatra <bogdan@kde.org>
This is consistent with other platforms such as iOS and
XCB, which return a QMimeData with an empty text when the clipboard
is empty.
[ChangeLog][Android] QClipboard::mimeData() now returns an empty
object instead of null pointer from QClipboard when clipboard is
empty for consistency with other platforms.
Change-Id: I17068f0afcb63690cf11048ffa60e19dc9b08691
Task-number: QTBUG-41817
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
The ownership of the object passed into QClipboard::setMimeData()
is documented to be transferred to the clipboard, but we never
deleted it, thus all these objects would leak.
[ChangeLog][Android] Fixed memory leak in QClipboard::setMimeData()
Change-Id: I43e6bad1071be5f56c219cb9341584edba54d2bd
Task-number: QTBUG-41852
Reviewed-by: BogDan Vatra <bogdan@kde.org>
QClipboard::clear() is implemented by calling
QPlatformClipboard::setMimeData() with a null pointer. Since we
would do nothing in this case on Android, then the clear()
function would have no effect.
[ChangeLog][Android] Added support for QClipboard::clear()
Task-number: QTBUG-41854
Change-Id: Id569b102f2e561e46967b52f89d9b54031d92456
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Do the cheap test before sending an expensive query that might
use a mutex.
Task-number: QTBUG-41369
Change-Id: I78f03c84e5bbf0492f1b7ea18d1baa752a1beff2
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
For QLatin1String, startsWith/endsWith is overloaded, so comparing to
a latin-1 (C) string literal is efficient, since strlen() is
comparatively fast.
OTOH, QStringLiteral, when not using RVO, litters the code with
QString dtor calls, which are not inline. Worse, absent lambdas,
it even allocates memory.
So, just use QLatin1String instead.
Change-Id: I7eaf464c67b55116f970fe7f58a85f19ce4d254e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
For QLatin1String, operator== is overloaded, so comparing to a latin-1
(C) string literal is efficient, since strlen() is comparatively fast.
OTOH, QStringLiteral, when not using RVO, litters the code with
QString dtor calls, which are not inline. Worse, absent lambdas,
it even allocates memory.
So, just compare using QLatin1String instead.
Change-Id: I7af3bf3a67c55dae33ffaf9922d004fa168a3f9c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
For const char*s, operator== is overloaded, so comparing to a (C) string
literal is efficient, since qstrcmp doesn't require the length of the
strings to compare.
OTOH, QByteArrayLiteral, when not using RVO, litters the code with
QByteArray dtor calls, which are not inline. Worse, absent lambdas,
it even allocates memory.
So, just compare with a (C) string literal instead.
Change-Id: Id3bfdc89558ba51911f6317a7a73c287f96e6f24
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Return a system palette that sets the correct
highligh (text selection) colors.
Change-Id: Ic0a18f931913f28bdc73d9ec4ae347fe38839f17
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
[NSView setWantsBestResulutionOpenGLSurface] is a
hint and the driver may ignore it, for example when
using software OpenGL on a virtual machine.
In these cases devicePixelRatio() must return a value
corresponding to the actual OpenGL surface size. Use
[NSView convertSizeToBacking] which is one of the
recommended methods.
Task-number: QTBUG-41767
Change-Id: Ia79242219908a2454a83b44b6eb7463372764162
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
The code shifts "1" by a uint32 number. If that number is >= 32,
we're triggering undefined behavior, and Coverity rightfully complains
about that. Add some asserts to silence those warnings.
Change-Id: Ib91085a279b0a2b7ad37afad05ec1d764c0496b1
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
In cb8445f032 we removed the historical
+1 in font heights and the corresponding -1 in the descent measurement
for all font engines. But the change to the direct write font engine
was lost at some point during the transition to QPA, where we seem
to have integrated an older version of the source file.
[ChangeLog][Windows][Fonts] Fix off-by-one in font descent when using
the DirectWrite font engine.
Task-number: QTBUG-41783
Change-Id: Iffa24b5f2b4f6cc3a1f0034fdff63a1ee62ea9f7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
We were missing brackets, but luckily the only result was that we
unconditionally flushed events through flushWindowSystemEvents.
Change-Id: If10bcc6a07501b9fb0db891e01b8ecc9d794ab30
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
When the focus object changes to an object that does not have
IM enabled, we should hide the keyboard instead of resetting it.
This happens, for instance, if you change to a different tab in
an application while a text input in the previous tab had focus.
Previously the input panel would stay open and overlap part of the
new tab.
[ChangeLog][Android] Hide input panel when focus changes to an
object that does not accept input.
Change-Id: I4009d203559582fa02c15fdc1a3f009396af2531
Task-number: QTBUG-32399
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
It doesn't belong in QIOScreen, and simplifies the flow when changing
the focus window or the window state of the focus window. Both will
result in calling updateProperties on the view-controller, which will
re-configure the statusbar visibility and hide/show it as appropriate,
before triggering a re-layout of its own view, which will in turn
trigger an update of the screen properties based on the new statusbar
state, before re-layouting of QWindow based on the new screen state.
Change-Id: I89077a3fb5f843949ce833e4e727d2c753ea2eb6
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Add missing transformation for first rectangle of a region
consisting of multiple rectangles.
Task-number: QTBUG-41300
Change-Id: I1a25c422c93ceade79de79d5c49891e0a7211943
Reviewed-by: MihailNaydenov <garfieldhq@yahoo.com>
Reviewed-by: Alessandro Portale <alessandro.portale@digia.com>
Print a warning when the developer/user has explicitly
requested the software backend, but loading openglsw.dll fails.
Change-Id: I8b1f079538ac70ceacbf1a397a276659c760a4cc
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
The code which extracts style assets for the Android style
is licensed under the Apache license, which is not compatible
with LGPLv2.1. It is, however, compatible with LGPLv3. This
means that the Android platform plugin cannot be LGPLv2.1
as long as this code is included.
To minimize licensing confusion, we default to only providing
LGPLv3 for Android. If you want to build a LGPLv2.1-compatible
library, you can add -no-android-style-assets to the
configuration. This will in turn enable the LGPLv2.1 in
the configure output, and it will disable the extraction
code in the platform plugin.
Running the Android style with an LGPLv2.1-compatible platform
plugin will work, but it will look horrible.
[ChangeLog][Android] Default open-source license for
Qt for Android is now LGPLv3. For compatibility with the LGPLv2.1
license, add "-no-android-style-assets" to your configuration.
Change-Id: I6c7b52140f38138520871fa7c69debbb4ee90e6c
Task-number: QTBUG-41365
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
On API-11+ we are going to use PopupMenu instead of ContextMenu to show
context menus. A PopupMenu displays a Menu in a modal popup window
anchored to a View. The popup will appear below the anchor view if there
is room, or above it if there is not.
Task-number: QTBUG-39736
Change-Id: Ie412ab0935b868348ce5c8bb0bf53571ffefd582
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Also fixes the handling of shareContext() for contexts and format() for windows
and makes QOffscreenSurface working.
Change-Id: I3c3374a9de14a5b8428de3e11d9d7e1285c5b9c7
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Windows accessibility can crash if it handles an event with an invalid
QAccessibleInterface pointer (one whose object has been deleted).
Task-number: QTBUG-41597
Change-Id: Iba099f7cb732fd00f18f04bd951c6cdd94785871
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
A fullscreen window can have maximized flags being set so first check
the NetWmStateFullScreen flag.
Change-Id: Ia802abf3cfa4c784baa2d55088e3f53310f0362e
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
This reverts commit 636d2e3402.
The issue was caused by a bug in ANGLE, not a lack of hardware support.
Change-Id: If2a66cd023dc7f2329dc2812169042487eecd428
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@digia.com>
Tor Arne Vestbø
Paul Olav Tvete
Richard Moe Gustavsen
Dyami Caliri
Gabriel de Dietrich
Friedemann Kleint
Alexander Volkov
Jan Arve Saether
Morten Johan Sørvig
Frederik Gladhorn
Lorn Potter
Marc Mutz
Timur Pocheptsov
Eskil Abrahamsen Blomfeldt
Alessandro Portale
Jake Petroules
Oswald Buddenhagen
J-P Nurmi
Giuseppe D'Angelo
Tor Arne Vestbø
Eskil Abrahamsen Blomfeldt
Kai Koehne
BogDan Vatra
Gabriel de Dietrich
Laszlo Agocs
Nico Vertriest
Andrew Knight